Catal Lett (2009) 129:250–257
DOI 10.1007/s10562-008-9810-0
Characterization and Catalytic Properties of the Ni/Al2O3
Catalysts for Aqueous-phase Reforming of Glucose
Guodong Wen Æ Yunpeng Xu Æ Zhusheng Xu Æ
Zhijian Tian
Received: 2 November 2008 / Accepted: 29 November 2008 / Published online: 13 December 2008
Ó Springer Science+Business Media, LLC 2008
Abstract Two Ni/Al2O3 catalysts with high Ni loadings
of 36 wt% and 48 wt%, which possess high activities for
aqueous-phase reforming of glucose, have been successfully prepared by a novel two-step impregnation method.
The catalytic performance was investigated at 533 K and
autogenous pressure in a batch reactor and a significant
enhancement in hydrogen yield was observed over the
catalyst prepared by two-step impregnation as compared to
the corresponding catalyst prepared by conventional single
impregnation. The catalysts were characterized by XRD,
N2 adsorption/desorption, H2 chemisorption, O2 uptake,
TPO, H2-TPR and TEM. It was found that two-step
impregnation yielded catalysts with higher nickel dispersion as well as smaller nickel particle size compared to
single impregnation.
Keywords Aqueous-phase reforming Glucose Hydrogen Ni/Al2O3
1 Introduction
Biomass was considered as a promising renewable source
for hydrogen production compared to the diminishing fossil
fuels [1]. Recently, a new process has been reported by
Dumesic and co-workers that hydrogen could be efficiently
produced from biomass-derived glucose and polyols at
mild conditions near 500 K in a single fixed-bed flow
reactor aqueous-phase reforming (APR) process [2]. This
process is energy saving because part of the water and the
oxygenated hydrocarbons with low volatility were not
allowed to vaporize. Moreover, APR occurs at low temperatures that the water–gas shift (WGS) reaction is
thermodynamically favored, thus lead to low levels of CO.
Particular attention had been paid to glucose [3–6], the
only compound in this class of reactants that is directly
relevant to biomass utilization. APR of glucose occurs
according to the following stoichiometric reaction [7]:
C6 H12 O6 þ 6H2 O ! 12H2 þ 6CO2
ð1Þ
The following two main reactions would occur during
the reaction conditions:
G. Wen Y. Xu Z. Xu Z. Tian (&)
Laboratory of Applied Catalysis, Dalian Institute of Chemical
Physics, Chinese Academy of Sciences, 457 Zhongshan Road,
Dalian 116023, People’s Republic of China
e-mail: [email protected]
G. Wen
Graduate University of Chinese Academy of Sciences,
19 Yuquan Road, Beijing 100049, People’s Republic of China
Z. Tian
State Key Laboratory of Catalysis, Dalian Institute of Chemical
Physics, Chinese Academy of Sciences, 457 Zhongshan Road,
Dalian 116023, People’s Republic of China
123
C6 H12 O6 ! 6H2 + 6CO
ð2Þ
CO þ H2 O ! CO2 þ H2
ð3Þ
Pt catalysts were identified as promising catalysts for the
APR reactions [7]. However, the high cost of Pt makes it
advantageous to develop non-precious catalysts. Although
non-precious Ni catalyst showed high initial activity that
was comparable to Pt catalyst, significant deactivation was
observed [8]. Efforts have been made to improve the
catalytic activities of the Ni catalysts by impregnating
another metal components [9, 10]. Furthermore, modified
skeletal Ni catalysts were developed to improve the
Ni/Al2O3 Catalysts for Aqueous-phase Reforming of Glucose
stability of the aforementioned bimetallic catalysts
[11, 12]. In the present study, a novel two-step
impregnation technique for the preparation of the Ni/
Al2O3 catalysts was developed, and the catalysts were
tested in the APR of glucose. It was noteworthy that the
catalysts with high Ni loadings prepared by two-step
impregnation exhibited significantly superior catalytic
properties than the corresponding catalysts prepared by
conventional single impregnation.
2 Experimental
2.1 Catalyst Preparation
The catalysts were prepared by impregnation Al2O3
with a solution containing the required amount of
Ni(NO3)26H2O. After impregnating with Ni, the samples
were evaporated at 353 K to dryness, then dried at 383 K
overnight, and finally calcined in air at 723 K for 3 h. The
Al2O3 used was derived from pseudo-boehmite. It has been
calcined in air at 773 K for 3 h prior to impregnation. Two
groups of Ni catalysts were prepared according to the
following detailed steps.
1.
2.
Single impregnation. The calcined as-prepared samples with different Ni loadings of 14 wt%, 21 wt%,
27 wt%, 36 wt%, and 48 wt% were denoted as
14 wt% Ni/Al2O3(S), 21 wt% Ni/Al2O3(S), 27 wt%
Ni/Al2O3(S), 36 wt% Ni/Al2O3(S), and 48 wt%
Ni/Al2O3(S), respectively.
Two-step impregnation. In the case of two-step impregnation, the first impregnation was carried out by
impregnation, evaporation, drying and calcination,
and then the same operation with the same additive
amount of Ni was repeated. The calcined as-prepared
catalysts with Ni loadings of 27 wt%, 36 wt%, and
48 wt% were denoted as 27 wt% Ni/Al2O3(T), 36 wt%
Ni/Al2O3(T), and 48 wt% Ni/Al2O3(T), respectively.
Prior to catalyst ex situ studies such as XRD, N2
adsorption/desorption, and TEM, all of the Ni/Al2O3 catalysts were reduced in flowing H2 for 2 h at 823 K, then
cooled to room temperature and passivated in a stream of
O2/N2 (1% O2).
2.2 Catalyst Characterization
The chemical compositions (Ni loadings) of the samples
were determined with a Philips Magix601 X-ray fluorescence (XRF) analysis apparatus. Powder XRD patterns of
the catalysts were recorded on a PANalytical X’Pert PRO
diffractometer using Cu Ka radiation (k = 1.5404 Å) at
40 kV and 40 mA. A NOVA-4000 physical adsorption
251
instrument was used to measure the N2 adsorption–
desorption isotherms of the samples at 77 K. The specific
surface area was determined from the linear portion of the
BET plot. The pore-size distribution was calculated from
the desorption isotherm using the Barrett–Joyner–Halenda
(BJH) formula. Prior to the measurements, the samples
were degassed in vacuum at 623 K for 2 h to remove
physically adsorbed components. Transmission electron
microscopy (TEM) measurements were carried out on a
FEI Tecnai G2 microscope. The samples were ultrasonically suspended in ethanol and placed onto a copper grid.
H2 pulse chemisorption was measured at room temperature in a homemade setup with a TCD detector. A catalyst
sample of 100 mg was reduced in H2 flow at 823 K for 2 h.
After reduction, the temperature was kept in a Ar flow at
823 K for 2 h and thereafter allowed to cool to room
temperature for in situ chemisorption measurement. The
exposed Ni atoms were assumed as H/Ni atomic ratio of
1:1. The oxygen pulse uptake was measured at 723 K with
4.5% O2/Ar. The amount of oxygen consumed was detected using an online mass spectrometry (OminiStar GSD
300, Balzers).
Temperature-programmed reductions (TPR) of catalysts
were conducted with 5% H2/Ar mixed gas at a flow rate of
37 mL/min (STP) from 303 to 1,073 K at a constant
heating rate (10 K/min). Temperature-programmed oxidation (TPO) of catalyst was conducted with 4.5% O2/Ar
mixed gas at a flow rate of 37 mL/min (STP) from 303 to
1,073 K at a constant heating rate (10 K/min). The gas
effluent was monitored online with a mass spectrometry
mentioned above.
2.3 Catalytic Activity Test
The APR of glucose was performed in a homemade batch
reactor described elsewhere [5]. The reaction was carried
out under autogenous pressure with continuous magnetic
stirring. In a typical experiment, 1.0 g catalyst, 90 g H2O
and the required amount of C6H12O6H2O were put inside
the reactor. The reactor was thoroughly purged with argon,
and then heated to the desired temperature at a certain
heating rate. The reaction time started after reaching the
desired temperature. When the reaction finished, the reaction was quenched by cold water. The quantity of the gas
product was measured after the reaction by wet gas flowmeter. Prior to the activity test, the Ni catalysts were
reduced for 2 h at 823 K and ambient pressure with a
heating rate of 2 K/min.
The gas products were analyzed according to the method
described elsewhere [13]. The liquid product was drained
after reaction, and then analyzed by inductively coupled
plasma atomic emission spectrometry (ICP-AES), gas
chromatography-mass spectroscopy (GC6890-MS5973 N,
123
252
G. Wen et al.
Agilent) with a HP-Innowax capillary column, HPLC
(Elite, RI detector, Sugar SC-1011 column (Shodex)) and
total organic carbon (TOC).
3 Results and Discussion
3.1 Physicochemical Characterization
3.1.1 BET Surface Area and Pore Structure
Table 1 showed the physical properties of support and
Ni/Al2O3 catalysts. All samples exhibited type IV isotherms typical of a mesoporous texture. As the Ni loadings
of the Ni/Al2O3 catalysts prepared by single impregnations
increased from 14 to 48 wt%, the surface area (SBET) and
the pore volume (Vp) gradually decreased, which suggested
a negative effect of Ni on texture properties. However, the
pore diameter of the catalysts did not change significantly.
Moreover, 36 wt% Ni/Al2O3(T) and 48 wt% Ni/Al2O3(T)
showed smaller surface area and pore volume than 36 wt%
Ni/Al2O3(S) and 48 wt% Ni/Al2O3(S), respectively,
whereas 27 wt% Ni/Al2O3(T) exhibited larger surface area
and pore volume than 27 wt% Ni/Al2O3(S).
3.1.2 X-ray Diffraction (XRD)
Figure 1 showed the XRD patterns of the samples calcined
at 723 K for 3 h, and the characteristic peaks of NiO were
marked in the figure. As expected, the characteristic peaks
of NiO were gradually intensified, and their peak width at
half height lessened as the Ni content increased for the
catalysts prepared by single impregnations. The results
indicated that the amount of detectable crystalline NiO and
the particle size of NiO increased with increasing of the Ni
content. It can be also seen from the figure that the characteristic peaks of NiO for 27 wt% Ni/Al2O3(T), 36 wt%
Table 1 Physical properties of support and Ni/Al2O3 catalysts
Catalysts
SBET (m2/g)
Vp (cm3/g)a
Dp (nm)b
Al2O3
249.7
0.54
6.60
14 wt% Ni/Al2O3(S)
195.7
0.45
6.56
21 wt% Ni/Al2O3(S)
187.9
0.41
6.53
27 wt% Ni/Al2O3(S)
181.0
0.38
6.59
36 wt% Ni/Al2O3(S)
159.3
0.31
6.52
48 wt% Ni/Al2O3(S)
124.4
0.27
6.62
27 wt% Ni/Al2O3(T)
184.1
0.45
6.57
36 wt% Ni/Al2O3(T)
138.3
0.30
6.56
48 wt% Ni/Al2O3(T)
116.2
0.25
6.57
Fig. 1 Powder XRD spectra of Ni/Al2O3 samples calcined at 723 K
for 3 h
Ni/Al2O3(T) and 48 wt% Ni/Al2O3(T) samples were
weaker and broader than those for 27 wt% Ni/Al2O3(S),
36 wt% Ni/Al2O3(S) and 48 wt% Ni/Al2O3(S) samples,
respectively. The results suggested that the calcined Ni/
Al2O3 catalysts prepared by two-step impregnations had
smaller amount of detectable crystalline NiO as well as
smaller NiO particle size than the corresponding catalysts
prepared by single impregnations. In addition, three broad
peaks at 2h = 37.5°, 46°, and 67° in the XRD curve of bare
alumina indicated the presence of amorphous c-Al2O3.
The XRD patterns of reduced Ni/Al2O3 catalysts were
shown in Fig. 2. Decreasing Ni content in the catalysts prepared by single impregnations resulted in broader and weaker
characteristic peaks of Ni0. In addition, the characteristic
peaks of Ni0 for 27 wt% Ni/Al2O3(T), 36 wt% Ni/Al2O3(T)
and 48 wt% Ni/Al2O3(T) samples were broader and weaker
than those for 27 wt% Ni/Al2O3(S), 36 wt% Ni/Al2O3(S) and
48 wt% Ni/Al2O3(S), respectively. The results indicated that
the amount of detectable crystalline Ni0 and the particle size
of metallic Ni increased with increasing of the Ni content for
the catalysts prepared by single impregnations, while the
catalysts prepared by two-step impregnations had smaller
amount of detectable crystalline Ni0 and metallic Ni particle
size than the corresponding catalysts prepared by single
impregnations. Furthermore, the NiO and metallic Ni particle
size were calculated by Scherre equation from NiO(200) and
Ni(111) planes, respectively, the results agreed well with the
above deduction.
3.1.3 Transmission Electron Microscopy (TEM)
a
Pore volume is calculated from the desorption branch of physisorption isotherm
b
BJH desorption average pore diameter
123
The micrographs of reduced 36 wt% Ni/Al2O3 and 48 wt%
Ni/Al2O3 catalysts prepared by single and two-step
Ni/Al2O3 Catalysts for Aqueous-phase Reforming of Glucose
253
Ni/Al2O3(S) and 48 wt% Ni/Al2O3(S) were found to yield
larger Ni particle size compared to 36 wt% Ni/Al2O3(T)
and 48 wt% Ni/Al2O3(T), respectively. The above results
from TEM are in accordance with the results from XRD.
3.1.4 H2 Temperature Programmed Reduction (H2-TPR)
Fig. 2 Powder XRD spectra of Ni/Al2O3 samples reduced at 823 K
for 2 h
impregnations are shown in Fig. 3. It can be seen from the
figures that Ni particles were irregularly dispersed on
platelet-like Al2O3 support. In addition, the 36 wt%
The H2-TPR curve of calcined Ni/Al2O3 samples was
shown in Fig. 4. Figure 4a exhibited the TPR patterns of
Ni/Al2O3 catalysts prepared by single impregnations. For
the 27 wt% Ni/Al2O3(S) sample, three peaks were
observed: a high temperature peak at ca. 1,060 K and two
low temperature peaks at ca. 545 K and ca. 800 K,
respectively. The two low temperature peaks were attributed to reduction of NiO with weaker interaction with
Al2O3 such as amorphous NiO and crystalline NiO,
respectively, while the high temperature peak was attributed to reduction of NiAl2O4 [14, 15]. The three peaks
were difficult to discern as the Ni loadings decreased from
27 to 14 wt%. When the Ni loadings increased from 27 to
48 wt%, the high temperature peak shifted towards lower
temperature, indicating that higher metal content favored
Fig. 3 TEM images of
Ni/Al2O3 catalysts reduced
at 823 K for 2 h.
a 36 wt% Ni/Al2O3(S),
b 36 wt% Ni/Al2O3(T),
c 48 wt% Ni/Al2O3(S),
and d 48 wt% Ni/Al2O3(T)
123
254
G. Wen et al.
Fig. 4 H2-TPR curves of
Ni/Al2O3 samples calcined
at 723 K for 3 h. a Catalysts
prepared by single impregnation
with Ni loadings ranging from
14 to 48 wt%. b 27 wt%
Ni/Al2O3, 36 wt% Ni/Al2O3
and 48 wt% Ni/Al2O3 catalysts
prepared by single
impregnations and two-step
impregnations
the formation of more readily reducible Ni species [16, 17].
Moreover, a low temperature peak at ca. 700 K with high
intensity was observed, whereas the low temperature peak
at ca. 545 K severely attenuated. The results indicated
that large amount of crystalline NiO was formed, whereas
the amount of amorphous NiO remarkably decreased.
Figure 4b compared the H2-TPR profiles of 27 wt%
Ni/Al2O3, 36 wt% Ni/Al2O3 and 48 wt% Ni/Al2O3 catalysts prepared by single impregnations and two-step
impregnations. For the 27 wt% Ni/Al2O3 samples, only
one high temperature peak at ca. 1,020 K was observed on
27 wt% Ni/Al2O3(T), whereas two additional lower temperature peaks were found on 27 wt% Ni/Al2O3(S). For the
36 wt% Ni/Al2O3 samples, only two peaks at ca. 860 K
and 1,040 K were shown on 36 wt% Ni/Al2O3(T), whereas
a additional lower temperature peak at ca. 700 K was
exhibited on 36 wt% Ni/Al2O3(S). For the 48 wt%
Ni/Al2O3 samples, all of the three peaks which were similar to those for 36 wt% Ni/Al2O3(S) were displayed on
both the 48 wt% Ni/Al2O3(S) and 48 wt% Ni/Al2O3(T)
samples. However, the three peaks shifted towards slightly
higher temperatures for the 48 wt% Ni/Al2O3(T) sample.
The results indicated that the less reducible Ni species with
strong interaction with the support were more prevalent in
catalysts prepared by two-step impregnations, and was
associated with the formation of smaller Ni particles as
shown in Table 2.
3.1.5 H2 Chemisorption and O2 Uptake
Based on the assumption that unreduced NiO is present in a
separate dispersed phase in intimate contact with the support, we calculated the Ni dispersion from the combination
of the H2 chemisorption and O2 uptake data [18, 19]. As
listed in Table 2, the Ni surface area exhibited a maximum
at Ni loading of 27 wt% for the samples prepared by single
impregnations. The 36 wt% Ni/Al2O3(T) and 48 wt%
Ni/Al2O3(T) catalysts showed higher Ni surface area than
36 wt% Ni/Al2O3(S) and 48 wt% Ni/Al2O3(S), respectively, whereas 27 wt% Ni/Al2O3(T) exhibited lower Ni
Table 2 Results of H2 chemisorption, O2 uptake and XRD tests for Ni/Al2O3
Catalysts
Chemisorptiona
Ni surface area
(SANi, m2 gcat-1)
XRDb
Reduction degree
(f, %)
Ni dispersion
(DNi, %)
14 wt% Ni/Al2O3(S)
4.0
24.6
17.2
21 wt% Ni/Al2O3(S)
10.3
48.4
27 wt% Ni/Al2O3(S)
36 wt% Ni/Al2O3(S)
11.4
9.3
39.6
45.3
48 wt% Ni/Al2O3(S)
7.9
27 wt% Ni/Al2O3(T)
6.5
36 wt% Ni/Al2O3(T)
48 wt% Ni/Al2O3(T)
dNiO (nm)
dNi (nm)
–
2.9
15.4
5.0
3.0
16.2
8.0
11.0
28.7
6.9
26.3
51.8
4.7
30.0
38.5
18.3
20.1
5.6
4.0
11.9
41.1
12.1
12.4
7.4
13.4
53.5
7.8
24.3
20.0
a
DNi (%) = 1.17X/(Wf), where X is the H2 uptake from H2 chemisorption, W is the weight percent of Ni in sample, f is the reduction degree
obtained from O2 uptake data and DNi is the Ni dispersion corrected for the amount reduced to metal [18, 19]
b
dNiO and dNi are from XRD data of NiO (200) and Ni (111) planes, respectively
123
Ni/Al2O3 Catalysts for Aqueous-phase Reforming of Glucose
255
surface area than 27 wt% Ni/Al2O3(S). Higher Ni dispersions were observed on 27 wt% Ni/Al2O3(T), 36 wt%
Ni/Al2O3(T) and 48 wt% Ni/Al2O3(T) as compared to
27 wt% Ni/Al2O3(S), 36 wt% Ni/Al2O3(S) and 48 wt%
Ni/Al2O3(S), respectively. The results indicated that twostep impregnations effectively promoted the dispersion of
Ni compared to single impregnations.
from 14 to 48 wt%, and the catalysts prepared by two-step
impregnations exhibited smaller metallic Ni particle size
than the corresponding catalysts prepared by single
impregnations. Combing the results from Table 3, it can be
found that high hydrogen selectivities were observed over
catalysts with metallic Ni particle size larger than ca. 7 nm
for the catalysts prepared by single impregnations. In a
parallel, similar high hydrogen selectivities were observed
over 36 wt% Ni/Al2O3(T) and 48 wt% Ni/Al2O3(T) with
metallic Ni particle size larger than 7 nm, whereas low
hydrogen selectivity was obtained on 27 wt% Ni/Al2O3(T)
with smaller metallic Ni particle size. It was assumed that
the catalytic properties of the catalysts were related to the
metallic Ni particle size. It seems that the APR of glucose
over these Ni catalysts is a structure sensitive process,
analogous to the APR of glycerol on Pt/Al2O3 catalysts [20].
Although the conversion of glucose for 48 wt%
Ni/Al2O3(T) reached nearly 95%, only 32.6% of carbon in
the initial feed was converted into the gas products, indicating that large amount of carbon was either converted
into the liquid phase or deposited on the catalyst. Propanal,
acetone, 2-butanone, ethanol, 1-propanol, cyclopentanone,
1-pentanol, 3-pentanol, 2,5-hexanedione, 2,3-dimethylcyclopentenone, 5-ethyldihydro-2-furanone, and phenol
were found by GC-MS in the liquid effluent. In addition,
the carbon balance based on TOC analysis of the liquid
product showed that 34.1% of carbon in the initial feed
deposited on the catalyst. The carbon deposition was further identified by TPO characterization. In a parallel, it was
3.2 Catalytic Activity
3.2.1 Effect of Ni Content and Impregnation Method
on Catalytic Performance
Table 3 showed the results for the APR of glucose over the
Ni/Al2O3 samples. It can be seen from the table that the
hydrogen selectivities and yields increased gradually from
Ni loadings ranging from 14 to 21 wt%, then greatly
increased to 27 wt%, and finally showed slight increase in
the range of 27–48 wt%. It can be also seen from the table
that 36 wt% Ni/Al2O3(T) and 48 wt% Ni/Al2O3(T) gave
higher conversion of C to gas, hydrogen yields and methane selectivities, as well as lower CO molar concentrations
than 36 wt% Ni/Al2O3(S) and 48 wt% Ni/Al2O3(S),
respectively, whereas 27 wt% Ni/Al2O3(T) exhibited lower
hydrogen yield, hydrogen selectivity, and CO concentration than 27 wt% Ni/Al2O3(S).
It can be seen from Table 2 that the metallic Ni particle
size of the samples prepared by single impregnations
increased from 2.9 to 38.5 nm as the Ni loadings increased
Table 3 Experimental results for the APR of 2.4 wt% glucose at 533 K and autogenous pressure over Ni/Al2O3 for 4 h
14 wt% Ni/ 21 wt% Ni/ 27 wt% Ni/ 36 wt% Ni/ 48 wt% Ni/ 27 wt% Ni/ 36 wt% Ni/ 48 wt% Ni/ Blanka
Al2O3(S)
Al2O3(S)
Al2O3(S)
Al2O3(S)
Al2O3(T)
Al2O3(T)
Al2O3(T)
Al2O3(S)
Catalysts
Glucose conversion (%) 100
Conversion of C
7.6
to gas (%)
100
7.8
100
12.6
100
13.1
94.7
16.5
100
12.6
95.2
33.3
94.8
32.6
96.3
6.0
Gas-phase composition
H2 (mol.%)
18.1
24.3
40.4
41.7
46.9
38.0
47.3
48.5
12.3
CO2 (mol.%)
75.4
69.2
54.2
53.6
49.0
57.1
46.5
45.2
80.2
CO (mol.%)
4.9
5.0
4.0
3.3
2.3
3.4
0.3
0.3
7.0
CH4 (mol.%)
0.2
0.5
1.0
0.9
1.4
0.9
5.3
5.4
0.1
C2–C6 hydrocarbons
(mol.%)
H2 selectivity (%)b
1.5
0.9
0.5
0.5
0.4
0.6
0.6
0.5
0.4
10.5
15.6
33.3
35.2
43.5
30.1
44.0
46.3
7.0
CH4 selectivity (%)
0.2
0.7
1.6
1.5
2.6
1.4
9.9
10.4
0.2
Alkane selectivity (%)c
6.4
4.7
4.1
4.0
4.8
4.3
12.8
13.0
1.6
0.8
1.2
4.2
4.6
7.2
3.8
14.7
15.1
0.4
H2 yield (%)
a
d
Experiment performed in the absence of metal catalyst
b
H2 selectivity (%) = (moles of H2 produced/moles of C atoms in gas phase) 9 2 9 100 [2]
c
Alkane selectivity (%) = (moles of C in gaseous alkanes/total moles of C in gas phase) 9 100 [2]
d
Hydrogen yield is defined as the amount of hydrogen in the gaseous products normalized by the amount of hydrogen that would be present if
the carbon atoms in the initial feed had all participated in the reforming reaction
123
256
found that part of the carbon was reversibly adsorbed on
the catalyst and could be washed away by ethanol, which
was consistent with the reference describe elsewhere [21].
The above results indicated that the side degradation
reactions severely occurred during the APR of glucose
[22]. However, it has been reported recently by Dumesic
et al. that the degradation liquid products could be further
efficiently converted to specific liquid fuels by integrating
several flow reactors operated in a cascade mode [23].
3.2.2 Effect of Residence Time on Catalytic
Performance
The effect of residence time on catalytic performance was
shown in Fig. 5. It can be seen from the figure that the
hydrogen selectivities exhibited a maximum at ca. 60 min.
The molar concentrations of CO decreased with the
increasing of the residence time within the considered
residence time range. The hydrogen yields increased
sharply in the residence time range of 2–60 min, then
gradually increased to 360 min, and finally decreased
slowly after 360 min.
It was reported that high concentrations of glucose lead
to low hydrogen selectivities because of the homogeneous
decomposition reactions [2], thus low hydrogen selectivity
was observed at the initial stage of the reaction due to the
high glucose initial concentration. In addition, it can be
seen from Table 3 that the homogeneous degradation
reactions lead to high CO concentration as shown in the
blank experiment. Therefore, it can be deduced that the
homogeneous degradation reactions were severely promoted at the initial stage of the reaction, thus lead to high
CO concentration. It seems that the reforming reaction
gradually dominated with the decrease of glucose concentrations in the residence time range of 60–240 min, thus
lead to high hydrogen selectivities as well as low CO
Fig. 5 Effect of residence time on the APR of 2.4 wt% glucose at
533 K and autogenous pressure over 48 wt% Ni/Al2O3(T)
123
G. Wen et al.
concentrations. However, the methanation reaction seems
gradually to be dominant at longer residence time, thus
lead to low hydrogen selectivity.
3.3 Catalytic Stability
To investigate the catalytic stability, the 48 wt% Ni/
Al2O3(T) sample was used in repeated run. After the first
run, the catalyst was separated by centrifugation and
washed with ethanol and water, fresh glucose and water
were added, and the mixture was used for the next run.
However, the glucose conversion decreased to 75.2% after
the first run. In addition, the hydrogen yield decreased to
7.2%, which was equivalent to only 47.7% of the yield for
the fresh catalyst. The metallic Ni particle size of the spent
catalyst from the XRD patterns increased to 23.8 nm
compared to the metallic Ni particle size of 20.0 nm for the
fresh catalyst, indicating that the Ni catalyst slightly sintered during the APR reaction conditions. The TPR profiles
of the spent and fresh calcined 48 wt% Ni/Al2O3(T) were
shown in Fig. 6. It can be seen from the figure that the
spent catalyst took up hydrogen below 823 K at which
temperature it was reduced. It can be also seen that the
spent catalyst exhibited lower temperature peaks of H2
consumption than the fresh calcined catalyst. The results
indicated that the reduced catalyst were partly oxidized
during APR reaction conditions, and the interaction
between metal species and support was weakened during
the APR reaction conditions. Moreover, inductively coupled plasma atomic emission spectrometry (ICP-AES)
analysis of the liquid products showed a Ni effluent concentration of 38 wppm after the first run.
Fig. 6 H2-TPR profiles of fresh calcined and spent 48 wt% Ni/
Al2O3(T) catalysts
Ni/Al2O3 Catalysts for Aqueous-phase Reforming of Glucose
4 Conclusions
Two-step impregnations were found to yield catalysts with
higher Ni dispersions compared to single impregnations,
thus probably lead to smaller metal particle size and
stronger interaction between metal species and support.
Significantly higher hydrogen yields were observed over
catalysts with high Ni content prepared by two-step
impregnations such as 36 wt% Ni/Al2O3(T) and 48 wt%
Ni/Al2O3(T) as compared to the corresponding catalysts
prepared by single impregnations, whereas lower hydrogen
yield was obtained over catalysts with low Ni content
prepared by two-step impregnations such as 27 wt%
Ni/Al2O3(T).
Acknowledgments We thank Prof. Hengyong Xu and Mr. Xiangang
Ma for their help with HPLC analysis. We also thank Mr. Yingchong
Ma and Ms. Yang Yang for their help with TOC analysis.
References
1. Ni M, Leung DYC, Leung MKH, Sumathy K (2006) Fuel Proc
Technol 87:461
2. Cortright RD, Davda RR, Dumesic JA (2002) Nature 418:964
3. Davda RR, Dumesic JA (2004) Chem Commun 36
4. Tanksale A, Wong Y, Beltramini JN, Lu GQ (2007) Int J
Hydrogen Energ 32:717
257
5. Xu YP, Tian ZJ, Wen GD, Xu ZS, Qu W, Lin LW (2006) Chem
Lett 35:216
6. Valenzuela MB, Jones CW, Agrawal PK (2006) Energy Fuel
20:1744
7. Davda RR, Shabaker JW, Huber GW, Cortright RD, Dumesic JA
(2005) Appl Catal B: Environ 56:171
8. Davda RR, Shabaker JW, Huber GW, Cortright RD, Dumesic JA
(2003) Appl Catal B: Environ 43:13
9. Shabaker JW, Huber GW, Dumesic JA (2004) J Catal 222:180
10. Iriondo A, Barrio VL, Cambra JF, Arias PL, Güemez MB,
Navarro RM, Sánchez- Sánchez MC, Fierro JLG (2008) Top
Catal 49:46
11. Huber GW, Shabaker JW, Dumesic JA (2003) Science 300:2075
12. Shabaker JW, Huber GW, Dumesic JA (2004) J Catal 222:180
13. Wen GD, Xu YP, Ma HJ, Xu ZS, Tian ZJ (2008) Int J Hydrogen
Energ 33:6657
14. Vos B, Poels E, Bliek A (2001) J Catal 198:77
15. Zhu XL, Huo PP, Zhang YP, Cheng DG, Liu CJ (2008) Appl
Catal B: Environ 81:132
16. Zhang J, Xu HY, Jin XL, Ge QJ, Li WZ (2005) Appl Catal A:
Gen 290:87
17. Alberton AL, Souza MMVM, Schmal M (2007) Catal Today
123:257
18. Bartholomew CH, Pannell RB (1980) J Catal 65:390
19. Mustard DG, Bartholomew CH (1981) J Catal 67:186
20. Lehnert K, Claus P (2008) Catal Commun 9:2543
21. Wang ZX, Pan Y, Dong T, Zhu XF, Kan T, Yuan LX, Torimoto
Y, Sadakata M, Li QX (2007) Appl Catal A: Gen 320:24
22. Minowa T, Fang Z, Ogi T, Várhegyi G (1998) J Chem Eng Jpn
31:131
23. Kunkes EL, Simonetti DA, West RM, Serrano-Ruiz JC, Gärtner
CA, Dumesic JA (2008) Science 322:417
123